OOGUN / A CENSUS OF THE FDA DRUG SHORTAGE LIST

Still waiting

On 2026-09-01, the FDA's official drug shortage list held 67 medicines in active shortage. The median one has been there four and a half years. Four have been there since 2012. This page is that list, drawn in full, and a caution about the one mistake everyone makes when reading it.

Kenny Obidele · data pulled 2026-09-01 from the FDA's openFDA shortages endpoint · every number on this page is computed in the open

67drugs in active shortage

4.5 yrsmedian time on the list

58with a presentation fully unavailable

75%injectables

4on the list since 2012

The list, in full

Each line is one drug, from the day it was posted to the wall of the present. Nothing here has ended; every line is still growing. Hover any line for the drug behind it. Thick lines have at least one presentation fully unavailable.

The emergency room is the epicentre

This is not a list of exotic medicines. Anaesthesia, analgesia and the cardiovascular drips dominate: fentanyl, morphine, lidocaine, atropine, epinephrine combinations, dopamine, dobutamine. The drugs of the operating theatre and the crash cart. 75% of the list is injectable, which is the tell: sterile injectables are low-margin, hard to manufacture, and concentrated in a handful of plants, so when a line goes down there is nowhere for the volume to go.

The bulge of 16 entries from 2023 is the wave you heard about: the ADHD-stimulant and chemotherapy shortages. Three years on, they are still here.

The one mistake everyone makes with this list

You cannot read the ages on a shortage list as if they were typical shortage lengths. A snapshot is biased in a precise, correctable way, and it is the same mathematics as waiting for a bus.

A shortage that lasts ten years sits on the list for ten years of days. One that resolves in three months gives you only three months to catch it. Look on any single day and you over-sample the long ones, in exact proportion to their length. The identity: the mean duration among the shortages you catch is E[X²]/E[X], not E[X], and the two agree only when durations never vary.

In the simulation above the truth is known: a median shortage of 1 year. The snapshot reads 4.3 years, a 3.9x exaggeration of the mean, from nothing but the act of looking on one day.

So is the real list exaggerating too? Partly, but not enough to explain itself away. Renewal theory lets the inversion run backwards: under a steady-state assumption, the ages observed on 2026-09-01 are most consistent with typical durations around 9 years (90% interval 8 to 11). That assumption is genuinely violated (16 of the 67 arrived in 2023 alone), so treat the exact number as a model-based reading. What survives every caveat is the direction: a list whose median entry has already waited 4.5 years cannot be produced by shortages that mostly resolve in months. The stuck ones are not the exception. They are the list's character.

Why they stay stuck

Where a reason is stated at all, the two biggest are a demand increase and the manufacturer simply discontinuing the product. Both are the economics of generic injectables: prices too low to justify new capacity, so when demand rises or a plant fails, no one steps in. The list also carries 157 separate discontinuation notices, drugs on the way out entirely, which this census counts apart from the 67 active shortages.